Engine-fed Kilonovae (Mergernovae) -- II. Radiation

TL;DR

This paper models the radiation from engine-fed kilonovae powered by long-lived magnetars, predicting early blue bumps and late brightening, with spectral evolution slower than typical kilonovae, but overall luminosity similar due to limited heating efficiency.

Contribution

It combines dynamical wind-ejecta interaction with shock-heating and X-ray irradiation to model photon diffusion and predict observable signatures of engine-fed kilonovae.

Findings

Early blue bump in light curves within 1 day

Late-stage brightening if magnetar does not spin down

Spectral energy distribution peaks in blue early and shifts red more slowly

Abstract

The radioactive power generated by materials within the ejecta of a binary-neutron-star (BNS) merger powers an optical transient known as a kilonova. When the central remnant of a BNS merger is a long-lived magnetar, it continuously produces a highly magnetized wind, altering both the dynamics and temperature of the ejecta, leading to the expected emergence of an engine-fed kilonova. In the first paper of this series, we conducted a detailed study of the dynamics of wind-ejecta interaction and the efficiency of energy injection through shocks. In this work, we combine this dynamical evolution with both shock-heating and additional X-ray irradiation to model photon diffusion within a constant-opacity ejecta. By calculating the radiation, we obtain the light curve and spectral energy distribution (SED). Our findings reveal that, with energy injection, a blue bump typically appears in the early stages ($\lesssim 1$ day). Furthermore, if the magnetar has not spun down by that time, a brightening in the later stages occurs. Despite this, in a large parameter space, the expected luminosity of the engine-fed kilonova is not significantly higher than the typical r-process kilonova due to limited heating efficiency. The SED of engine-fed kilonovae peaks in the relatively blue band in the early stages and evolves towards the red, but at a slower rate compared to the typical r-process kilonova.